Average Trivia Enjoyer (ATE)

This code is intended to store a server application that provides the user answers to trivia-like questions. It is part of a computer science assignment (2nd year, 1st semester).

(Self imposed) deadline: 29 / 12 / 2022

Hard deadline: 19 / 01 / 2023

This project will be available on Github once the hard deadline passes.

License

Once the hard deadline passes, this project will be available under the MIT license. The code and its documentation were kindly contributed by:

• Maria-Miruna ALDICA (@miruna15aldica) - The GUI module;
• Matei-Ștefan IONESCU (@Matei77) - The Server module;
• Valentin-Ioan VINTILĂ (@w1bb) - The AI module.

Running the project

There are three ways to run this project, and each comes with its advantages and disadvantages:

• Running the project locally without venv - this has the benefit of running out of the box as long as a few requirements are met (please read the dedicated subsection);
• Running the project locally with venv - this is almost the same as the first option, only it is safer, but it consumes more time;
• Running the project using docker - this will take even longer, but will work on any machine.

Pick your poison!

Running locally without venv

In order to run this project, a few assumptions are made:

• The user is on a Linux OS (preferably Ubuntu or Arch);
• The user has make, python3.10 and pip3 installed, alongside a browser to check out the resulting server.

To build and run the project, please follow these steps:

1. From the parent (current) folder, run make build. This will create a folder called product that will contain all the files required by the server. The command will also check to see if the required packages are installed and, if they are missing, it will install them. If you dislike this behaviour, check out the other ways of running the server.
2. Run make run in order to start the server. In a browser (e.g. Firefox), you can now go to 127.0.0.1:8080 and check the functionality of the application. Please note that the terminal in which you typed the command will now act as a logger. You should also make sure the 8080 port is not already in use.
3. When you are done, you can exit the server by pressing Ctrl+C.
4. (Optional) Once you are done testing the functionality, you can run a cleanup procedure by typing make clean. Please make sure the server is not running in the background before running this command.

Running locally with venv

In order to run this project, a few assumptions are made:

• The user is on a Linux OS (preferably Ubuntu or Arch);
• The user has make, python3.10, pip3 and venv installed, alongside a browser to check out the resulting server. Note that venv should come with the Python package, but this is not guaranteed (Debian might not ship python with venv).

To build and run the project, please follow these steps:

1. From the parent (current) folder, run make venv-build. This will create a folder called venv-product that will contain all the files required by the server. The command will also install the required packages in the virtual environment.
2. Run make venv-run in order to start the server. In a browser (e.g. Firefox), you can now go to 127.0.0.1:8080 and check the functionality of the application. Please note that the terminal in which you typed the command will now act as a logger. You should also make sure the 8080 port is not already in use.
3. When you are done, you can exit the server by pressing Ctrl+C.
4. (Optional) Once you are done testing the functionality, you can run a cleanup procedure by typing make venv-clean. Please make sure the server is not running in the background before running this command.

Running using docker

In order to run this project, a few assumptions are made:

• The user is on a Linux OS (preferably Ubuntu or Arch);
• The user has docker correctly installed and configured on its machine. This might require the user to get added to a special group.

To build and run the project, please follow these steps:

1. From the parent (current) folder, run make docker-build. This will create the files inside docker-product and then will prompt docker to create an image file. This process will take a while.
2. After the first step is complete, one should get a message similar to this one:

   (...)
   Step 13/13 : CMD ["make", "run-here"]
   ---> Running in bcdf1f83fd11
   Removing intermediate container bcdf1f83fd11
   ---> ee0ed09054f9
   Successfully built ee0ed09054f9
   Successfully tagged w1bb/ia4:1.0
   
   Everything was setup successfully!
   

   In this case, the resulting image is called ee0ed09054f9 (it can be determined as it follows the Successfully built message). Use that name to start the actual server - this can be accomplished by running:

   wi@arch:~$ docker run -p 8080:8080 <IMAGE NAME>
   

   Based on the example above, one should run docker run -p 8080:8080 ee0ed09054f9. In a browser (e.g. Firefox), you can now go to 127.0.0.1:8080 and check the functionality of the application. Please note that the terminal in which you typed the command will now act as a logger. You should also make sure the 8080 port is not already in use.
3. When you are done, you can exit the server by pressing Ctrl+C.
4. (Optional) Once you are done testing the functionality, you can run a cleanup procedure by typing make docker-clean. Please make sure the server is not running in the background before running this command.

Documentation

This section will present the inner-workings of the code.

Required packages

In order to function properly, the project requires a number of packages, each with its specific functionality:

• bs4, required by the AI module - it helps scrape the Wikipedia articles;
• flask, required by the Server module - it provides the functionality to instantiate the actual server;
• google, required by the AI module - it searches Google for the best Wikipedia articles;
• lxml, required by the AI module - it is used by BeautifulSoup;
• rich - used throught the project to make text beautiful;
• tensorflow, required by the AI module - it is used by the transformers package;
• transformers, required by the AI module - it is the actual brain of the operation;
• waitress, required by the Server module - it provides the functionality to host the server on localhost:8080.

Please note that all these packages can be installed via pip3.

The AI module

The AI folder contains the following files:

ai.py = The file containing the AI implementation;
ai-test.py = An example that tests the AI implementation by allowing an agent to
             ask as many questions as he desires (break with Ctrl+C).

ai.py

This file contains two important functions:

• ask_question(original_question)

  This function receives a question (original_question) and provides an answer in the form of a string. If no results could be found in a given time span, the "Sorry, could not find any answer on Wikipedia!" string is returned instead.

• get_urls(original_question, stop_no, urls)

  This function searches Google for a question (original_question) and tries to find stop_no answers; the result is returned in results (this design choice was made so that ask_question could call it as a separate process in order to limit its maximum time consumption).

The former function is created to be exported, whilst the latter is used inside ask_question and should not be relevant to the other parts of the project.

Note: The ask_question function is slow at times, please be patient!

ai-test.py

Run this script and try questions such as Who is the current president of the United States? or When did the roman empire fall?. Were the results satisfactory?

The GUI module

The GUI folder contains the following files:

static/ = The folder containing the styles and other static elements (in this
|         case, only the styles);
|-- styles/ = The folder containing the styles;
|   |-> style-answer.css = The CSS corresponding to the answer.html page;
|   |-> style-index.css = The CSS corresponding to the index.html page;
templates/ = The folder containing the HTML files;
|-> answer.html = The page loaded once an answer is provided by the bot;
|-> index.html = The query page.

These files contains the design of the program.

The server module

The web application is implemented using the Flask framework. When the user first loads the web page the server will render the index.html page. After a question is asked and submitted the server will be notified that the POST method was used and will get the question from the form and try to find the best answer for it, which will be given by rendering the answer.html page. The server is run using the waitress WSGI (Web Server Gateway Interface). The serve function starts the server and runs the application.